Turk J Urol 2020; 46(Supp. 1): S27-S39 • DOI: 10.5152/tud.2020.

20117
S27
ENDOUROLOGY
Invited Review

Artificial intelligence (AI) in urology-Current use and future

directions: An iTRUE study
Milap Shah1,2 , Nithesh Naik2,3 , Bhaskar K. Somani2,4 , BM Zeeshan Hameed1,2,5

Cite this article as: Shah M, Naik N, Somani BK, Hameed BMZ. Artificial intelligence (AI) in urology-Current use and future directions: An
iTRUE study. Turk J Urol 2020; 46(Supp. 1): S27-S39.

ORCID iDs of the authors:
M.S. 0000-0003-1668-7006;
N.N. 0000-0003-0356-7697;
B.K.S. 0000-0002-6248-6478;
B.M.Z.H. 0000-0002-2904-351X
1
Department of Urology,
Kasturba Medical College
Manipal, Manipal Academy of
Higher Education, Manipal,
Karnataka, India
2
i-TRUE: International Training
and Research in Uro-oncology
and Endourology, Manipal,
Karnataka, India
3
Department of Mechanical and
Manufacturing Engineering,
Manipal Institute of
Technology, Manipal Academy
of Higher Education, Manipal,
Karnataka, India
4
Department of Urological
Surgery, University Hospital
Southampton NHS Trust,
Southampton, UK
5
KMC Innovation Centre,
Manipal Academy of Higher
Education, Manipal, Karnataka,
India
Submitted:
09.04.2020
Accepted:
12.04.2020
Available Online Date:
27.05.2020
Corresponding Author:
BM Zeeshan Hameed
E-mail:
zeeshanhameedbm@gmail.com
©Copyright 2020 by Turkish
Association of Urology
Available online at
www.turkishjournalofurology.com
ABSTRACT
Objective: Artificial intelligence (AI) is used in various urological conditions such as urolithiasis, pediatric
urology, urogynecology, benign prostate hyperplasia (BPH), renal transplant, and uro-oncology. The various
models of AI and its application in urology subspecialties are reviewed and discussed.
Material and methods: Search strategy was adapted to identify and review the literature pertaining to the
application of AI in urology using the keywords “urology,” “artificial intelligence,” “machine learning,”
“deep learning,” “artificial neural networks,” “computer vision,” and “natural language processing” were
included and categorized. Review articles, editorial comments, and non-urologic studies were excluded.
Results: The article reviewed 47 articles that reported characteristics and implementation of AI in urologi-
cal cancer. In all cases with benign conditions, artificial intelligence was used to predict outcomes of the
surgical procedure. In urolithiasis, it was used to predict stone composition, whereas in pediatric urology
and BPH, it was applied to predict the severity of condition. In cases with malignant conditions, it was ap-
plied to predict the treatment response, survival, prognosis, and recurrence on the basis of the genomic and
biomarker studies. These results were also found to be statistically better than routine approaches. Appli-
cation of radiomics in classification and nuclear grading of renal masses, cystoscopic diagnosis of bladder
cancers, predicting Gleason score, and magnetic resonance imaging with computer-assisted diagnosis for
prostate cancers are few applications of AI that have been studied extensively.
Conclusions: In the near future, we will see a shift in the clinical paradigm as AI applications will find their
place in the guidelines and revolutionize the decision-making process.
Keywords: Artificial intelligence; deep learning; machine learning; prostate cancer; urolithiasis; urology.
Introduction data, the future health care system is likely to
move toward AI outpatient clinics and preven-
Artificial intelligence (AI) refers to the compu- tive medicine. AI provides more accuracy and
tational capability of the machine to mimic and reliable clinical decisions; hence, it is possibly
perform human cognitive tasks. It is causing going to be an integral part of the health care
a paradigm shift in terms of providing health system.
care and decision-making for the clinicians.
The advances in the medical technologies The four subfields of AI in health care are as
used in health care, such as electronic medical follows:
records (EMRs), are providing humongous
amounts of data.[1] This large amount of data 1. Machine learning (ML): ML is statisti-
allows computer-based predictions and deci- cal technique-based programming that allows
sions to be made to aid in better patient care a computer system to learn and recognize
(Figure 1). By 2025, the growth rate of AI patterns to model without explicit instruc-
applications in health care is expected to be tions. ML uses procedural computer programs
29.3%, and the global revenue is estimated to wherein machines are trained to learn, detect
increase by 40%.[2] With the available patient data patterns, compute, and infer from the

S28
Figure 1. A schematic process chart for building an artificial intelligence model
datasets provided. It is observed that the machines are able
to generate results similar to those generated by human intel-
ligence.
2. Natural language processing (NLP): It illustrates the ability
of a computer to comprehend the written and spoken language.
Some applications that are possible through NLP include lan-
guage translation, text processing, and speech recognition. In
order to extract useful information and reliable details from
patient services and provide “virtual assistance” for physi-
cians, a comprehensive data research such as electronic medical
record (EMR), doctor’s notes, pharmaceutical products, and
medical imaging can also be analyzed.
3. Deep learning (DL) and artificial neural networks (ANNs):
In the network architectural layers, the ANN comprises of indi-
vidual units that function like artificial neurons programmed
to accomplish computer tasks and recognize complex patterns.
DL requires training massive datasets of multilayered neural
networks. Deep neural convolution network (DCNN) is a com-
monly used ANN, which is effective when used in digitized
image pattern identification or recognition.
Main Points:
• AI is widely used in the diagnosis, treatment, and outcome pre-
diction in various urological conditions.
• In urolithiasis, AI is used to detect stone composition and to
predict spontaneous passage of stone.
• AI applications in prostate carcinoma are used for the diag-
nosis-Gleason scoring, treatment decisions-making, and even
predicting the disease-free survival.
Turk J Urol 2020; 46(Supp. 1): S27-S39
DOI: 10.5152/tud.2020.20117
4. Computer vision: Computer vision technology is used for
visual search, trend forecasting, augmented reality, and virtual
reality. The radiological and pathological images and simple
and complex endoscopic videos can be used by machines to
understand the details and patterns in the images in order to
identify the tumors or malignancy present in the diagnostic
images. The latest experience at human level in diagnostic
imaging has already shown that AI has extensive “knowledge”
to identify tumors. Computer vision can also be used for analy-
sis and grading of pathological tissue slides.
AI is increasingly applied not only to the diagnosis of urologi-
cal conditions but also its management and predictive analysis.
[3]
This article focuses on addressing the application of AI and
AI algorithms in urological subspecialties. This article reviews
the use of AI in various benign and malignant conditions such as
urolithiasis, pediatric urology, urogynecology, benign enlarge-
ment of the prostate, renal transplant, and uro-oncology pertain-
ing to the kidneys, bladder, prostate, and testes.
Clinical and Research Consequences
Application of AI in Benign Urological Conditions
Urolithiasis
In the past few decades, there has been a rapid transition in the
analysis, treatment and monitoring of cases with urolithiasis, the
recent entry being applications of AI to identify the stone from
computed tomography (CT) and ultrasound (US) images[4,5],
detecting stone composition[6,7], predicting spontaneous stone
passage[8,9] and even the outcomes of endourological procedures
(Table 1).[10-13]
Shah et al. Artificial intelligence in urology S29

Table 1. Studies looking at applications of AI in benign urological conditions (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

A. Urolithiasis

Parakh et al.[4] Urinary stone • 535 patients Convolutional >90% NA NA

detection on unen- (279 stones pre- neural network
hanced CT images sent; 256 stones (CNN)
absent)
• 100 scans (test
data)

Chiang et al.[5] To identify asso- • 151 (calcium • Discriminant • Genetic factors NA NA

ciation between oxalate stone) analysis DA: 64% ANN:
stone disease and patients • Artificial Neural 65 %
genetic polymorp- network (ANN) • Genetic and
• 105 healthy
hisms, patient Env. Factors
controls
habits DA: 75% ANN:
89%

Kazemi et al.[6] For early detecti- • 936 patients • -Bayesian 97.1% (ensemble NA NA
on of the type of dataset model-Decision model)
kidney stone and • 42 features Trees
the most influenti- • ANN
al parameters • Rule-based
classifiers

Kriegshauser To investigate use • 32 stone dataset -Multiparametric • 97% to distingu- NA NA

et al.[7] of single-source algorithms ish uric acid and
dual-energy com- non-uric acid
puted tomography stones
(ssDECT) for the • 72 % to distin-
characterization of guish non-uric
renal stones. acid stone
subtypes

Eken et al.[8] To compare AI • 227 patients • ANN NA ANN: 94.9% ANN: 78%
models with LR • 176 urinary • Genetic algo- GA: 67.6% GA: 76%
by applying on stones rithm (GA) LR: 95.5% LR: 48%
medical dataset • 51 no stones • Logistic regres-
sion analysis
(LR)

Dal Moro et al.[9] To predict the • 1163 patients Linear program- NA 84.5% 86.9%
spontaneous (402 found ming support
passage of ureteral valuable) vector machine
stones in patients • nine clinical (LPSVM)
with renal colic factors

Shabaniyan To predict posto- • 254 patients • Machine 94.8% Requirement of NA

et al.[10] perative outcome • 26 variables learning (ML) Stent placement:
of PCNL techniques such 85.2% Require-
as sequential ment of blood
forward selecti- transfusion: 95%
on and Fisher’s
discriminant
analysis
 Turk J Urol 2020; 46(Supp. 1): S27-S39
S30 DOI: 10.5152/tud.2020.20117

Table 1. Studies looking at applications of AI in benign urological conditions (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

Aminsharifi To predict mul- • 146 adult pati- • ML-based SVM 80%-95.1% Stone free status: NA
et al.[11] tiple outcomes ents model 92% Need for
after percutaneous repeat PCNL: 97%
nephrolithotomy Need for ESWL:
(PCNL) and 82% Need for
compare with URS: 91%
GSS and CROES
nomogram
Kadlec et al.[12] For outcome • 382 renal units • nonlinear LR Classification ac- Stone free status: Stone free status:
prediction after model curacy of 69.6% 75.3% Need for 60.4% Need for
various forms of sec. procedure: sec. procedure:
endourological 30% 98.3%
intervention
Seckiner et al.[13] To predict stone • 139 patients ANN 88.7% in the test NA NA
free status after (training set) group
ESWL • 32 patients
(validation set)
• 32 patients (test
set)
B. Benign Prostatic Hyperplasia
Torshizi et al.[14] To diagnose the • 44 patients Fuzzy system ~90% NA NA
severity of BPH expert
and suggest app-
ropriate treatment
Sonke et al.[15] To diagnose BPH • 1903 patients ANN NA 71% 69%
and compare with
regression analysis
C. Pediatric Urology
Bagli et al.[16] To predict sonog- • 84 children ANN 100% 100% 100%
raphic outcome training set
after pyeloplasty • 16 children test
in children with set
ureteropelvic junc-
tion obstruction
Logvinenko To predict patients • 2259 patients ANN Multivariate NA For any grade For any grade
et al.[17] at high risk of LR analysis VUR VUR
VCUG abnorma- • ANN: 64% • ANN: 60%
lities, based on • MLR: 84% • MLR: 25%
RBUS findings
Blum et al.[18] To predict the • 55 patients • Linear support 93% 91% 96%
need for surgery in • 45 features vector machine
UPJO cases based (SVM)
on dynamics of
renogram
D. Urogynecology
Sabadell et al.[20] To predict occur- • 169 patients ML algorithm NA NA NA
rence of SUI after
prolapse surgery
and as a diagnostic
tool
Shah et al. Artificial intelligence in urology S31

Table 1. Studies looking at applications of AI in benign urological conditions (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

Jelovsek et al.[21] To predict recur- • 1301 patients LR models NA NA NA

rence, complica-
tions, and health
status improve-
ment after prolapse
surgery
E. Renal Transplant
Atallah et al.[22] To predict 5-year • 2728 patients • Naïve Bayes Ba- 80.77% 81.2% NA
graft survival (70% training sed Feature Se-
and 30% testing) lector (NBBFS)
Algorithm
• K-nearest neigh-
bor Algorithm
(KNN)
Greco et al.[23] To predict graft fa- • 194 patients • ML algorithms NA 88.2% 73.8%
ilure and associa- • Decisional Trees
tion with BMI and
other risk factors
Goldfarb et al.[24] To predict cadave- 37,407 patients Logistic regression 65% NA NA
ric graft survival dataset based model Tree-
over three years based model
based on pretrans-
plant variables
CT: Computed Tomography; CNN: Convolutional Neural Network; ANN: Artificial Neural Network; DA: Discriminant Analysis; ssDECT: Single-Source Dual-Energy
Computed Tomography; LR: Logistic regression; GA: Genetic Algorithm; LP SVM: Linear Programming Support Vector Machine; PCNL: Percutaneous Nephrolithotomy;
ML: Machine Learning; GSS: Guys Stone Score; CROES: Clinical Research Office of the Endourological Society; SVM: Support Vector Machine; ESWL: Extracorporeal
Shock Wave Lithotripsy; URS: Ureterorenoscopy; BPH: Benign Prostatic Hyperplasia; VCUG: Voiding Cystourethrogram; RBUS: Renal Bladder Ultrasound; NBBFS:
Naïve Bayes Based Feature Selector; KNN: K-Nearest Neighbor; BMI: Body Mass Index; UPJO: Uretero-Pelvic Junction Obstruction

Parakh et al.[4] studied the diagnostic performance of the convo-
lution neural network (CNN) on CT images for detection of uri-
nary stones in 535 adult patients assumed to have renal calculi
using two scanners. The first scanner identified the urinary tract,
and the next one detected the stone. Using nine different varia-
tion models, it achieved an accuracy of more than 90%. The
study concluded that the efficiency of CNNs can be improved
by the use of transfer learning with datasets augmented with
labeled images. Shabaniyan et al.[10] developed a decision sup-
port system using ML techniques to predict the outcomes of
surgical treatment for renal calculus. The algorithm was trained
with a dataset of 254 patients and 26 parameters, which com-
prised variables from patients’ history, renal calculus composi-
tion, and laboratory investigations. This model achieved an
accuracy of 94.8%, 85.2%, and 95% in predicting outcomes
of a procedure, predicting whether patient will require a stent
after the procedure, and predicting the need for blood transfu-
sion, respectively. Aminsharifi et al.[11] studied data of 146 adult
patients in whom percutaneous nephrolithotomy (PCNL) was
done to validate efficiency of a machine-based learning algo-
rithm for predicting the outcomes after PCNL and to compare
the results with Clinical Research Office of Endourological
Society nomogram and Guy’s Stone Score (GSS). This program
predicted the PCNL results with an accuracy of up to 95%.
Benign Enlargement of the Prostate
Many questionnaires are available for the clinical prediction of
benign prostatic hyperplasia (BPH), yet the results are unreli-
able and inaccurate. Various AI techniques and ANN models
such as multilayered back propagation method to predict the
severity of obstruction on the basis of noninvasive tests have
been used (Table 1).[14,15] Torshizi et al.[14] applied fuzzy intel-
ligent systems in predicting the severity of BPH and also rec-
ommended the treatment required for it. The study consisted of
two models. The first model predicted the severity, whereas the
second model helped to make a treatment decision. The results
were then compared for accuracy and validation with an expert
panel. The accuracy achieved was nearly 90%.
Pediatric Urology
AI has been used in the field of pediatric urology for predicting
the outcome of surgical procedures[16], severity of the condi-
tion on the basis of imaging as well as detecting abnormalities
in imaging (Table 1).[17,18] Bagri et al.[16] applied computerized

S32
ANN to predict the outcomes after pyeloplasty on the basis of
US findings in children with uretero-pelvic junction obstruc-
tion. The prediction was based on whether postoperatively the
results were “significantly improved,” “improved,” “same,” or
“worse.” The results showed 100% sensitivity and specificity
for all four-outcome measures. Both multivariate analysis and
ML algorithms were used by Logvinenko et al.[17] to evaluate
whether the renal and bladder ultrasound (RBUS) could predict
the abnormalities on voiding cystourethrogram (VCUG) for
conditions such as vesico-ureteric reflux and congenital ure-
thral abnormalities. The results showed that RBUS was a poor
predictor of the abnormalities of VCUG and both could only
complement each other but cannot replace them.
Urogynecology
On the basis of urinary incontinence data obtained from the
wearable devices, AI techniques were applied to predict the
time and number of incontinence episodes and the outcome of
conservative or medical management for stress urinary inconti-
nence (SUI).[19] Models to predict the occurrence of complica-
tions such as SUI after prolapse surgery, recurrence, and overall
outcomes of surgery were also studied on the basis of the data
available from various randomized controlled trials (Table
1).[20,21] In the near future, AI applications can be used to provide
personalized care based on the patient demographics and clini-
cal characteristics of every individual.
Renal Transplant
The outcome of kidney transplant prediction is very important.
Various studies have been conducted to predict the outcomes
of kidney transplantation using ANN and ML algorithms, as
described (Table 1).[22-24] Atallah et al.[22] proposed a predic-
tion method by combining two methods-Bayes and k-nearest
neighbor-which achieved more accuracy by choosing minimum
number of features. It was based on data mining techniques to
predict five-year graft survival after transplantation. This new
proposed prediction method comprises three stages: data prepa-
ration stage, feature selection stage, and prediction stage. This
prediction method can be used in other transplant datasets to
measure the graft survival.
Application of AI in Uro-Oncology
Testicular Malignancy
Not much has been studied about the applications of AI in
testicular malignancy. Baessler et al.[25] applied ML-based CT
radiomics to determine whether the lymph nodes dissected in
patients with metastatic or advanced nonseminomatous testicu-
lar germ cell tumor were malignant or benign. The model cor-
rectly classified with an accuracy of 0.81 (area under the curve
[AUC]), 88% sensitivity, and 72% specificity.
Turk J Urol 2020; 46(Supp. 1): S27-S39
DOI: 10.5152/tud.2020.20117
Renal Cell Carcinoma
ML and DL algorithms based on CT-texture analysis were
applied for differentiating renal masses such as angiomyoli-
poma, clear cell renal cell carcinoma (ccRCC), papillary renal
cell carcinoma, and oncocytoma[26-28] to predict the nuclear
grade and to identify certain genetic mutations to predict the
prognosis, recurrence, and survival outcomes (Table 2). Kocak
et al.[26] used CT-texture analysis, applied ML techniques to
predict and identify the nuclear grade (Furhman) of ccRCC,
and compared the results with those obtained with percutane-
ous biopsy. The results were comparable, and the maximum
predictive value was achieved with the use of the support vec-
tor machine (SVM). The algorithm could differentiate nuclear
grades in 85.1% of ccRCC cases. Ding et al.[29] also conducted
a similar study showing increased precision in classifying the
grade of ccRCC.
Biomarkers and signatures based on more than one gene expres-
sion have been developed in recent years for predicting the
ccRCC overall survival (OS) and prognosis of the disease. Li et
al.[30] developed a model based on 15 genes, which could help
predict the prognosis and survival. They found that the group
with a higher risk had substantially poorer prognosis and sur-
vival than the group with patients having lower risk. The risk
groups were not associated with patient characteristics such as
sex or age but were related to hemoglobin levels. They were
also associated with tumor features such as size and grade.
PBRM1 mutations are the second most common mutations
found in ccRCC. Kocak et al.[31] applied ANN- and ML-based
algorithms to identify PBRM1 mutations based on CT scan
texture analysis. Overall, 88% of ccRCC with PBRM1 mutation
status was correctly identified by ANN. On the basis of these
results, future studies can be conducted to develop noninvasive
biomarkers for identifying histopathological subtypes to predict
the prognosis and response to treatment.
Bladder Cancer
ML algorithms, DCNN models, genetic algorithms, and SVMs
have been applied in bladder cancer for improving cystoscopic
diagnosis and prediction of prognosis and survival (Table 2).[32-
35]
Ikeda et al.[32] made a competent CNN by training it with
2102 cystoscopic pictures with an aim to increase the efficiency
in diagnosis of bladder cancer using AI. It achieved sensitivity
and specificity of 89.7% and 94.0%, respectively. Lorencin et
al.[33] used the data of 1997 and 986 images with and without
bladder cancer, respectively, to train multilayer perceptron
along with DCNN for the diagnosis of bladder malignancy. It
showed promising results, with AUC value reaching up to 0.99.
Wang et al.[36] achieved more than 75% accuracy by using least
squares SVM in predicting the five-year overall and cancer-
specific mortality of patients post radical cystectomy.
Shah et al. Artificial intelligence in urology S33

Table 2. Studies looking at application of AI in urological malignancies (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

A. Renal Cell Carcinoma(RCC)

Kocak et al.[26] To distinguish • 68 RCC pati- Artificial Neural ANN: 84.60% ANN: 69% ANN: 100%
the major subt- ents for inter- Network (ANN) SVM: 69% SVM: 71% SVM: 100%
ypes of RCC nal validation Support vector
• 26 RCC pati- machine (SVM)
ents for exter-
nal validation
• 275 CT images
for texture
features

Feng et al.[27] To differentiate • 58 patients • Machine 93.9% 87.8% 100%

angiomyolipoma • 42 features Learning (ML)
(AML) and RCC based quanti-
based on texture tative texture
analysis of CT analysis
images • SVM with re-
cursive feature
elimination
• Synthetic
minority
oversampling
technique
(SMOTE)

Coy et al.[28] To distinguish • 4000 iterations Deep Learning 74.4% 85.8% NA

ccRCC and on- (90% training (DL) based Go-
cocytoma from and 10% vali- ogle TensorFlow
MDCT images dation) software
• 179 patients

Ding et al.[29] To preoperatively • 92 cases • Logistic Reg- NA NA NA

distinguish high (for validation) ression (LR)
nuclear grade model
from low nuclear • Least absolute
grade in ccRCC shrinkage and
selection ope-
rator (LASSO)
for texture
score

Li et al.[30] To predict sur- N=533 (training • ML-based ran- NA NA NA

vival in patients dataset) Risk dom forest va-
with ccRCC score model riable hunting
based on gene based on 15 ge- Cox regression
expression nes N=101 (test analysis
dataset)

Kocak et al.[31] To identify the • 45 patients • ML-based ANN: 88.22% NA NA

mutation status (29 without quantitative RF: 95%
of PBRM1 gene mutation; 16 CT-texture
in ccRCC pati- with mutation) analysis such
ents • 161 labeled as
segmentations • Random Forest
(87 without (RF) algorithm
mutation; 74 • ANN
with mutation)
 Turk J Urol 2020; 46(Supp. 1): S27-S39
S34 DOI: 10.5152/tud.2020.20117

Table 2. Studies looking at application of AI in urological malignancies (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

B. Bladder Carcinoma

Ikeda et al.[32] To improve • 2102 images Convolutional NA 89.7% 94%

cystoscopic diag- (1671 normal Neural Network
nosis of bladder tissue; 431 (CNN)
cancer using AI tumor lesions)
• 8:2 (training:
test set)

Lorencin et al.[33] To use multilayer • 1997 bladder • Multilayer Per- NA NA NA

perceptron met- cancer images ceptron (MLP)
hod for diagnosis • 986 non- • Laplacian edge
of bladder cancer cancer tissues detector
from cystoscopic images
images

Hashemi et al.[34] To classify • 540 cystos- • Multilayer ne- 7% decrease NA NA

cystoscopic copic bladder ural networks in error on
bladder images images • Genetic algo- classification as
using AI rithm (GA) compared with
other methods

Eminaga et al.[35] To perform • 479 patients • Deep Learning F1 scores Xcep- NA NA

diagnostic clas- • 18,681 images CNN (DL- tion: 99.52%
sification based (generated CNN) ResNet: 99.48%
on cystoscopic with 10 degree • Xception
images using grades) model
DL-CNN • 60% training • ResNet50
set-10% vali- model
dation set • InceptionV3
• 30% test set • VGG-19
• VGG-16

Wang et al.[36] To predict blad- • 117 bladder • Output-based • 5 years overall • 5 years overall • 5 years overall
der cancer prog- cancer patients transfer lear- mortality mortality mortality
nosis in terms of ning approach Proposed Proposed Proposed
five-year overall with least classifier(v1): classifier(v1): classifier(v1):
and cancer-speci- square support 76.97% 78.48% 75.79%
fic mortality vector machine Proposed Proposed Proposed
(LS-SVM) classifier(v2): classifier(v2): classifier(v2):
76.18% • 78.29% 74.33% •
5-year cancer- • 5-year cancer- 5-year cancer-
specific morta- specific morta- specific morta-
lity Proposed lity Proposed lity Proposed
classifier(v1): classifier(v1): classifier(v1):
74.85% 90.26% 38% Proposed
Proposed Proposed classifier(v2):
classifier(v2): classifier(v2): 31%
75.15% 92.38%

Gavriel et al.[37] To predict five- • 78 patients • ML-based en- 94.8% 89.5% 97.4%
year prognosis of diagnosed with semble model
bladder cancer MIBC
Shah et al. Artificial intelligence in urology S35

Table 2. Studies looking at application of AI in urological malignancies (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

Hasnain et al.[38] To predict • Dataset of Ensemble ML- NA >70% >70%

postcystectomy 3503 patients based models
recurrence and • Support vector
survival machine
(SVM)
• K-nearest
neighbor Algo-
rithm (KNN)
• Random Forest
• Gradient-bo-
osted trees
(GBT)

Bartsch et al.[39] To predict recur- • 112 frozen • ML cased NA Test Set Five Test Set Five
rence of NMIBC NMIBC speci- genetic gene combined gene combined
based on genome mens programming rule: 69% rule: 62%
profile • 21 gene classi- algorithm Three gene com- Three gene com-
fier set bined rule: 71% bined rule: 67%

Wu et al.[40] To compare • 123 CT scans Multiple DL- Base DL-CNN Base DL-CNN Base DL-CNN
different DL- (pre and post- CNN models 70% 60% 80%
CNN models to treatment) with structure
predict response modification and
to treatment in layer freezing
bladder cancer
(T0 prediction)

C. Prostate Carcinoma

Ström et al.[41] To diagnose and Training set-976 ANN NA 99% 94.9%

grade prostate patients (6682
cancer in biop- slides) Test
sies set-246 patients
(1631 slides)

Bulten et al.[42] To assign Gle- 1243 patients DL system Benign versus Benign versus Benign versus
ason grade to (5759 biopsies) malignant: 96%- malignant: malignant: 83%-
prostate biopsies 97% 97.4% 100%
using AI Grade group 2 or Grade group 2 or Grade group 2 or
more: 79%-83% more: 86%-95% more: 52%-70%
Grade group 3 or Grade group 3 or Grade group 3 or
more: 76%-82% more: 76%-92% more: 72%-782%

Viswanath et To compare va- • 85 T2W MRI • Quadratic NA NA NA

al.[44] rious classifier in datasets Discriminant
detecting CaP on Analysis
t2W MRI images (QDA) -Sup-
using radiomic port Vector
texture features Machines
(SVMs)
• Naïve Bayes
Decision Trees
(NBDT)
 Turk J Urol 2020; 46(Supp. 1): S27-S39
S36 DOI: 10.5152/tud.2020.20117

Table 2. Studies looking at application of AI in urological malignancies (Continue)

Study Objective Study Design Algorithm/Model Accuracy Sensitivity Specificity

Wildeboer et For automated • 50 men with ML techniques NA NA NA

al.[45] localization of biopsy confir- using B-mode,
CaP based on med CaP shear-wave elas-
radiomics of tography (SWE),
TRUS and dynamic
contrast-enhan-
ced ultrasound
(DCE-US)
radiomics

Deng et al.[46] For treatment • 78 features ML-based model NA NA NA

stratification of associated with
patients with me- the patient cli-
tastatic castrate nical and me-
resistant CaP dical history,
lab reports and
metastases

de la Calle et To predict • 648 samples AI algorithm 100% in identifi- NA NA

al.[47] recurrence and (424 tumors, cation of ERG+
progression of 224 normal tumor
CaP based on bi- tissue)
omarker analysis • Tissue micro
assays anti
Ki-67, ERG
antibodies

Bibault et al.[48] To predict sur- • Dataset from AI algorithm 10-year OS: 87% 10-year OS: 60% NA
vival in patients PLCO trial 10-year cancer- 10-year cancer-
with CaP • 8776 patients specific survival: specific survival:
(diagnosed 98% 55%
with CaP on
follow-up)
• n=7021
(training set)
N=1755 (test
set)
RCC: Renal Cell Carcinoma; CT: Computed Tomography; ANN: Artificial Neural Network; SVM: Support Vector Machine; AML: Angiomyolipoma; ML: Machine Le-
arning; SMOTE: Synthetic Minority Oversampling Technique; ccRCC: Clear Cell Renal Cell Carcinoma; MDCT: Multiple Detector Computed Tomography; DL: Deep
Learning; LR: Logistic Regression; LASSO: Least Absolute Shrinkage and Selection Operator; PBRM1: Polybromo1; RF: Random Forest; CNN: Convolutional Neural
Network; MLP: Multi Layer Perceptron; AI: Artificial Intelligence; GA: Genetic Algorithm; DL-CNN: Deep Learning Convolutional Neural Network; LS-SVM: Least Squ-
are Support Vector Machine; MIBC: Muscle Invasive Bladder Cancer; KNN: K-Nearest Neighbor; GBT: Gradient-Boosted Trees; NMIBC: Non-Muscle Invasive Bladder
Cancer; CaP: Carcinoma Prostate; T2W MRI: T2 Weighted Magnetic Resonance Imaging; QDA: Quadratic Discriminant Analysis; NBDT: Naïve Bayes Decision Trees;
TRUS: Trans Rectal Ultrasonogram; SWE: Shear-Wave Elastography; DCE-US: Dynamic Contrast-Enhanced Ultrasound; PLCO: Prostate Lung Colorectal Ovarian; OS:
Overall Survival

Gavriel et al.[37] proposed an ensemble system comprising
ML-based algorithms to predict five-year prognosis with dif-
ferent combinations of image, clinical, and spatial features and
quantify potential prognostic markers related to lymphocytes,
macrophages, tumor buds, and PD-L1. The method success-
fully classified 71.4% of the patients who succumbed to muscle
invasive bladder cancer (MIBC) within five years, significantly
higher than the 28.6% of the current clinical gold standard, the
tumour, node, metastasis (TNM) staging system.
Several studies have applied ML-based algorithms and models
to identify genes that could predict the recurrence of disease or
the future progression. Slides of patients diagnosed with MIBC
were labeled with immunofluorescence (IF) and used for mea-
suring the tumor buds, to determine the effectiveness of neo-
adjuvant chemotherapy, and to identify patients who were not
responding to the treatment. This was done to stop the treatment
prematurely to avoid the adverse effects of chemotherapy.[38-40]
Shah et al. Artificial intelligence in urology
Prostate Carcinoma
AI applications are on the verge of revolutionizing the current
practice in carcinoma prostate (CaP) in terms of diagnosis, treat-
ment decisions, and even predicting the disease-free survival.
There is high observer-dependent variability in Gleason grading
because of the subjective nature of the analysis of biopsy speci-
mens. Considering this, Ström et al.[41] developed an AI model
for identification, Gleason grading, and localization of prostate
cancer. The model was trained with 6682 digitized slides of 976
men and tested on 1631 biopsy specimens from 246 men. It
achieved an accuracy of 0.997 (AUC) to differentiate between a
malignant and a benign tumor. The results in terms of Gleason
grading were also comparable to those achieved by the expert
pathologists. In various studies, DL methods to calculate the
Gleason Grading have been applied (Table 2).[42]
Multiparametric imaging uses multiple modalities or tech-
niques before making the ultimate diagnosis, and this adds
to the burden of the radiologist. However, in the current era,
computer-aided diagnosis is possible because of progress in
AI, which eventually helps in making the diagnosis by image
interpretation. This is particularly useful in situations where
multiple modalities, parameters, or techniques are involved in
diagnosing a condition.[43]
Application of a Quadrant Discriminant Classifier to the
radiomic features derived from T2-weighted MRI images
for detection of CaP[44] and application of ML-based random
Forrest classification algorithm to localize CaP on transrectal
ultrasonogram[45] have been studied (Table 2).
In view of the toxic effects of docetaxel chemotherapy, 20% of
the patients undergo therapeutic failure in metastatic castrate
resistant CaP. Deng et al.[46] developed an AI-based computa-
tional model that could differentiate patients in two groups,
docetaxel-tolerable and docetaxel-intolerable, for better and
individualized treatment for the patients in this category.
Identification of the presence of biomarkers on tissue microar-
rays can predict the risk of recurrence and metastasis. Biomarker
identification under IF microscope by the human eye is subjec-
tive as well as time-consuming. Hence, an automated method
using DL algorithms was developed for analysis of biomarkers
using 648 samples and IF staining with anti-Ki-67, ERG anti-
bodies. The results were promising, with only 5% difference
between manual and algorithm-based biomarker detection and
100% accuracy in identification of tumors positive for ERG.[47]
Bibault et al.[48] used data from the prospective clinical trial Prostate
Lung Colorectal and Ovarian cancer screening, selected patients
who were diagnosed with CaP during follow-up, and trained two
models to predict ten-year cancer-specific survival (CSS) and OS.
Of the 8776 patients diagnosed with PCa on follow-up, training of
S37
the models was done with 7021 and tested on dataset of 1755. It
achieved an accuracy of 0.87 and 0.98 for OS and CSS, respec-
tively. These models can be used online to provide predictions and
support informed decision-making in CaP treatment.
Limitations
AI applications are gaining significant interest in urology, but
their real-world implementation still faces an uphill task. There
are limitations to some studies that use AI algorithms and its
subsets in urological diseases. The key challenges that can be
addressed before being integrated into the clinical setting are the
incorporation of standardized criteria, the correction for system
variation, and the data collection from multiple institutions in
various geographical locations, so that the results can be gener-
alized and applied to the real-world scenario.[49]
Future Considerations
The President of the World Economic Forum, Klaus Schwabe,
made the following announcements at the Davos Summit just a
few years ago: “We stand on the brink of a technological revo-
lution that will fundamentally alter the way we live, work, and
relate to one another. In its scale, scope, and complexity, the
transformation will be unlike anything human kind has experi-
enced before”.[50]
Future work will concentrate on creating larger medical data-
bases and expanding AI techniques further. The use of enhanced
algorithms will take place on smartphones or can be accessed
through the cloud. Applications for clinical decision-making
and its use in the real world require appropriate permissions
from the regulatory bodies. Issues exist concerning the reliabil-
ity of a machine diagnosis and that prejudices of programming
do not create hindrances in the diagnosis.
Conclusion
In the near future, we will see a shift in the clinical paradigm
as AI applications will find their place in the guidelines and
revolutionize the decision-making process. Having said that,
human qualities of intelligence, adaptation, and sense of duty
will prove to be important factors in further development of AI.
Peer-review: This manuscript was prepared by the invitation of the
Editorial Board and its scientific evaluation was carried out by the
Editorial Board.
Author Contributions: Concept - M.S., N.N., B.M.Z.H.; Design
- N.N., B.M.Z.H.; Supervision - B.K.S.; Materials - M.S.; Data
Collection and/or Processing - M.S.; Analysis and/or Interpretation -
N.N., B.M.Z.H.; Literature Search - M.S., N.N., B.M.Z.H.; Writing
Manuscript - M.S., N.N., B.M.Z.H.; Critical Review - B.K.S.

S38
Conflict of Interest: The authors have no conflicts of interest to declare.
Financial Disclosure: The authors declared that this study has
received no financial support.
References
1. Beam A, Kohane I. Big Data and Machine Learning in Health
Care. JAMA 2018;319:1317. [Crossref]
2. Grand View Research. Healthcare Predictive Analytics Market
Analysis by Application (Operations Management, Financial, Pop-
ulation Health, Clinical), By End-Use (Payers, Providers), By Re-
gion (North America, Europe, Asia Pacific, Latin America, MEA)
and Segment Forecasts, 2018-2025. 2016 November 2016.
3. Drouin S, Yates D, Hupertan V, Cussenot O, Rouprêt M. A sys-
tematic review of the tools available for predicting survival and
managing patients with urothelial carcinomas of the bladder and of
the upper tract in a curative setting. World J Urol 2012;31:109-16.
[Crossref]
4. Parakh A, Lee H, Lee J, Eisner BH, Sahani DV, Do S. Urinary
Stone Detection on CT Images Using Deep Convolutional Neural
Networks: Evaluation of Model Performance and Generalization.
Radiology: Artificial Intelligence 2019;1:e180066. [Crossref]
5. Chiang D, Chiang H, Chen WC, Tsai FJ. Prediction of stone dis-
ease by discriminant analysis and artificial neural networks in
genetic polymorphisms: a new method. BJU Int 2003;91:661-6.
[Crossref]
6. Kazemi Y, Mirroshandel S. A novel method for predicting kidney
stone type using ensemble learning. Artif Intell Med 2018;84:117-
26. [Crossref]
7. Kriegshauser J, Silva A, Paden R, He M, Humphreys MR, Zell SI,
et al. Ex Vivo Renal Stone Characterization with Single-Source
Dual-Energy Computed Tomography: A Multiparametric Ap-
proach. Acad Radiol 2016;23:969-76. [Crossref]
8. Eken C, Bilge U, Kartal M, Eray O. Artificial neural network, ge-
netic algorithm, and logistic regression applications for predicting
renal colic in emergency settings. Int J Emerg Med 2009;2:99-105.
[Crossref]
9. Dal Moro F, Abate A, Lanckriet G, Arandjelovic G, Gasparella P,
Bassi P, et al. A novel approach for accurate prediction of sponta-
neous passage of ureteral stones: Support vector machines. Kidney
Int 2006;69:157-60. [Crossref]
10. Shabaniyan T, Parsaei H, Aminsharifi A, Movahedi MM, Jahromi
AT, Pouyesh S, et al. An artificial intelligence-based clinical de-
cision support system for large kidney stone treatment. Australas
Phys Eng Sci Med 2019;42:771-9. [Crossref]
11. Aminsharifi A, Irani D, Tayebi S, Kafash TJ, Shabanian T, Parsaei
H. Predicting the Postoperative Outcome of Percutaneous Neph-
rolithotomy with Machine Learning System: Software Validation
and Comparative Analysis with Guy's Stone Score and the CROES
Nomogram. J Endourol 2020;DOI: 10.1089/end.2019.0475.
[Crossref]
12. Kadlec A, Ohlander S, Hotaling J, Hannick J, Niederberger C, Turk
TM. Nonlinear logistic regression model for outcomes after endou-
rologic procedures: a novel predictor. Urolithiasis 2014;42:323-7.
[Crossref]
Turk J Urol 2020; 46(Supp. 1): S27-S39
DOI: 10.5152/tud.2020.20117
13. Seckiner I, Seckiner S, Sen H, Bayrak O, Dogan K, Erturhan S. A
neural network - based algorithm for predicting stone -free status
after ESWL therapy. Int Braz J Urol 2017;43:1110-4. [Crossref]
14. Torshizi A, Zarandi M, Torshizi G, Eghbali K. A hybrid fuzzy-
ontology based intelligent system to determine level of severity
and treatment recommendation for Benign Prostatic Hyperplasia.
Comput Methods Programs Biomed 2014;113:301-13. [Crossref]
15. Sonke GS, Heskes T, Verbeek AL, de la Rosette JJ, Kiemeney
LA. Prediction of bladder outlet obstruction in men with lower
urinary tract symptoms using artificial neural networks. J Urol
2000;163:300-5. [Crossref]
16. Bagli D, Agarwal S, Venkateswaran S, Shuckett B, Khoury AE,
Merguerian PA, et al. Artificial Neural Networks In Pediatric Urol-
ogy: Prediction Of Sonographic Outcome Following Pyeloplasty.
J Urol 1998;160:980-3. [Crossref]
17. Logvinenko T, Chow J, Nelson C. Predictive value of specific ul-
trasound findings when used as a screening test for abnormalities
on VCUG. J Pediatr Urol 2015;11:176.e1-176.e7. [Crossref]
18. Blum ES, Porras AR, Biggs E, Tabrizi PR, Sussman RD, Sprague
BM, et al. Early Detection of ureteropelvic junction obstruction
using signal analysis and machine learning: A dynamic solution to
a dynamic problem. J Urol 2018;199:847-52. [Crossref]
19. Bentaleb J, Larouche M. Innovative use of artificial intelligence
in urogynecology. Int Urogynecol J 2020; DOI: 10.1007/s00192-
020-04243-2. [Crossref]
20. Sabadell J, Salicrú S, Montero-Armengol A, Rodriguez-Mias N,
Gil-Moreno A, Poza JL. External validation of de novo stress uri-
nary incontinence prediction model after vaginal prolapse surgery.
Int Urogynecol J 2019;30:1719-23. [Crossref]
21. Jelovsek JE, Chagin K, Lukacz ES, Nolen TL, Shepherd JP, Barber
MD, et al. Models for predicting recurrence, complications, and
health status in women after pelvic organ prolapse surgery. Obstet
Gynecol 2018;132:298-309. [Crossref]
22. Atallah DM, Badawy M, El-Sayed A, Ghoneim MA. Predicting
kidney transplantation outcome based on hybrid feature selection
and KNN classifier. Multimed Tools Appl 2019;78:20383-407.
[Crossref]
23. Greco R, Papalia T, Lofaro D, Maestripieri S, Mancuso D, Bono-
figlio R. Decisional Trees in Renal Transplant Follow-up. Trans-
plant Proc 2010;42:1134-6. [Crossref]
24. Goldfarb-Rumyantzev AS, Scandling JD, Pappas L, Smout RJ,
Horn S. Prediction of 3-yr cadaveric graft survival based on
pre-transplant variables in a large national dataset. Clin Transpl
2003;17:485-97. [Crossref]
25. Baessler B, Nestler T, Pinto dos Santos D, Paffenholz P, Zeuch V,
Pfister D, et al. Radiomics allows for detection of benign and ma-
lignant histopathology in patients with metastatic testicular germ
cell tumors prior to post-chemotherapy retroperitoneal lymph node
dissection. Eur Radiol 2020;30:2334-45. [Crossref]
26. Kocak B, Yardimci A, Bektas C, Turkcanoglu MH, Erdim C, Yuce-
tas U, et al. Textural differences between renal cell carcinoma sub-
types: Machine learning-based quantitative computed tomography
texture analysis with independent external validation. Eur J Radiol
2018;107:149-57. [Crossref]
27. Feng Z, Rong P, Cao P, Zhou Q, Zhu W, Yan Z, et al. Machine
learning-based quantitative texture analysis of CT images of small
Shah et al. Artificial intelligence in urology
renal masses: Differentiation of angiomyolipoma without vis-
ible fat from renal cell carcinoma. Eur Radiol 2017;28:1625-33.
[Crossref]
28. Coy H, Hsieh K, Wu W, Nagarajan MB, Young JR, Douek ML,
et al. Deep learning and radiomics: the utility of Google Tensor-
Flow™ Inception in classifying clear cell renal cell carcinoma and
oncocytoma on multiphasic CT. Abdom Radiol 2019;44:2009-20.
[Crossref]
29. Ding J, Xing Z, Jiang Z, Chen J, Pan L, Qiu J, et al. CT-based
radiomic model predicts high grade of clear cell renal cell carci-
noma. Eur J Radiol 2018;103:51-6. [Crossref]
30. Li P, Ren H, Zhang Y, Zhou Z. Fifteen-gene expression based mod-
el predicts the survival of clear cell renal cell carcinoma. Medicine
2018;97:e11839. [Crossref]
31. Kocak B, Durmaz E, Ates E, Baykara Ulusan M. Radiogenom-
ics in Clear Cell Renal Cell Carcinoma: Machine Learning-Based
High-Dimensional Quantitative CT Texture Analysis in Predicting
PBRM1 Mutation Status. AJR Am J Roentgenol 2019;212:W55-
W63. [Crossref]
32. Ikeda A, Nosato H, Kochi Y, Kojima T, Kawai K, Sakanashi H,
et al. Support System of Cystoscopic Diagnosis for Bladder Can-
cer Based on Artificial Intelligence. J Endourol 2020;34:352-8.
[Crossref]
33. Lorencin I, Anđelić N, Španjol J, Car Z. Using multi-layer per-
ceptron with Laplacian edge detector for bladder cancer diagnosis.
Artif Intell Med 2020;102:101746. [Crossref]
34. Hashemia SMR, Hassanpourb H, Kozegarc E. Cystoscopic Image
Classification Based on Combining MLP and GA. Int J Nonlinear
Analys App 2020;11:93-105.
35. Eminaga O, Semjonow A, Breil B. Diagnostic classification of
cystoscopic images using deep convolutional neural networks. Eur
Urol Supp 2018;17:e1232. [Crossref]
36. Wang G, Zhang G, Choi K, Lam KM, Lu J. Output based transfer
learning with least squares support vector machine and its applica-
tion in bladder cancer prognosis. Neurocomputing 2020;387:279-
92. [Crossref]
37. Gavriel C, Dimitriou N, Brieu N, Nearchou IP, Arandjelovic
O, Schmidt G, et al. Identification of Immunological Features
Enables Survival Prediction of Muscle-Invasive Bladder Can-
cer Patients Using Machine Learning. BioRxiv 2020. DO:
10.1101/2020.02.24.963181 [Crossref]
38. Hasnain Z, Mason J, Gill K, Miranda G, Gill IS, Kuhn P, et al.
Machine learning models for predicting post-cystectomy re-
currence and survival in bladder cancer patients. PLoS One
2019;14:e0210976. [Crossref]
S39
39. Bartsch G, Mitra A, Mitra S, Almal AA, Steven KE, Skinner
DS, et al. Use of Artificial Intelligence and Machine Learning
Algorithms with Gene Expression Profiling to Predict Recurrent
Nonmuscle Invasive Urothelial Carcinoma of the Bladder. J Urol
2016;195:493-8. [Crossref]
40. Wu E, Hadjiiski LM, Samala RK, Chan HP, Cha KH, Richter C,
et al. Deep Learning Approach for Assessment of Bladder Cancer
Treatment Response. Tomography 2019;5:201-8. [Crossref]
41. Ström P, Kartasalo K, Olsson H, Solorzano L, Delahunt B, Ber-
ney DM, et al. Artificial intelligence for diagnosis and grading of
prostate cancer in biopsies: a population-based, diagnostic study.
Lancet Oncol 2020;21:222-32. [Crossref]
42. Bulten W, Pinckaers H, van Boven H, Vink R, de Bel T, van Gin-
neken B, et al. Automated deep-learning system for Gleason grad-
ing of prostate cancer using biopsies: a diagnostic study. Lancet
Oncol 2020;21:233-41. [Crossref]
43. Wildeboer R, van Sloun R, Wijkstra H, Mischi M. Artificial in-
telligence in multiparametric prostate cancer imaging with focus
on deep-learning methods. Comput Methods Programs Biomed
2020;189:105316. [Crossref]
44. Viswanath S, Chirra P, Yim M, Rofsky NM, Purysko AS, Rosen MA,
et al. Comparing radiomic classifiers and classifier ensembles for de-
tection of peripheral zone prostate tumors on T2-weighted MRI: a
multi-site study. BMC Med Imaging 2019;19:22. [Crossref]
45. Wildeboer R, Mannaerts C, van Sloun R, Budäus L, Tilki D, Wi-
jkstra H, et al. Automated multiparametric localization of prostate
cancer based on B-mode, shear-wave elastography, and contrast-
enhanced ultrasound radiomics. Eur Radiol 2019;30:806-15.
[Crossref]
46. Deng K, Li H, Guan Y. Treatment Stratification of Patients with
Metastatic Castration-Resistant Prostate Cancer by Machine
Learning. iScience 2020;23:100804. [Crossref]
47. de la Calle CM, Nguyen H, Hosseini-Asl E, So C, Socher R, Xiong
C, et al. Artificial intelligence for streamlined immunofluores-
cence-based biomarker discovery in prostate cancer. J Clin Oncol
2020;38(Suppl 6):279. [Crossref]
48. Bibault JE, Xing L. Predicting Survival in Prostate Cancer Pa-
tients with Interpretable Artificial Intelligence. Available at SSRN
3546050. 2020. [Crossref]
49. Chen J, Remulla D, Nguyen J, Aastha D, Liu Y, Dasgupta P, et al.
Current status of artificial intelligence applications in urology and
their potential to influence clinical practice. BJU Int 2019;124:567-
77. [Crossref]
50. Schwab, K. The Fourth Industrial Revolution; World Economic
Forum: Geneva, Switzerland, 2016; ISBN 9781944835002.